Fibrocystin/Polyductin Modulates Renal Tubular Formation by Regulating Polycystin-2 Expression and Function

BASIC RESEARCH www.jasn.org

ʈ Ingyu Kim,* Yulong Fu,† Kwokyin Hui,‡ Gilbert Moeckel,§ Weiyi Mai,* Cunxi Li,*¶ Dan Liang,* Ping Zhao,† Jie Ma,† Xing-Zhen Chen,** Alfred L. George, Jr.,* Robert J. Coffey,*¶ Zhong-Ping Feng,‡ and Guanqing Wu*†¶

Departments of *Medicine, §Pathology, and ¶Cell and Developmental Biology, Vanderbilt University, Nashville, Tennessee; †Division of Translational Cancer Research and Therapy, Cancer Hospital and Institute, Chinese Academy of Medical Sciences, Beijing, China; ‡Department of Physiology, University of Toronto, Toronto, Ontario, ʈ Canada; **Department of Physiology, University of Alberta, Edmonton, Alberta, Canada; and Department of Internal Medicine, First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China

ABSTRACT Autosomal recessive polycystic kidney disease is caused by mutations in PKHD1, which encodes the membrane-associated receptor-like protein fibrocystin/polyductin (FPC). FPC associates with the primary cilia of epithelial cells and co-localizes with the Pkd2 gene product polycystin-2 (PC2), suggesting that these two proteins may function in a common molecular pathway. For investigation of this, a mouse BASIC RESEARCH model with a gene-targeted mutation in Pkhd1 that recapitulates phenotypic characteristics of human autosomal recessive polycystic kidney disease was produced. The absence of FPC is associated with aberrant ciliogenesis in the kidneys of Pkhd1-deficient mice. It was found that the COOH-terminus of FPC and the NH2-terminus of PC2 interact and that lack of FPC reduced PC2 expression but not vice versa, suggesting that PC2 may function immediately downstream of FPC in vivo. PC2-channel activities were dysregulated in cultured renal epithelial cells derived from Pkhd1 mutant mice, further supporting that both cystoproteins function in a common pathway. In addition, mice with mutations in both Pkhd1 and Pkd2 had a more severe renal cystic phenotype than mice with single mutations, suggesting that FPC acts as a genetic modifier for disease severity in autosomal dominant polycystic kidney disease that results from Pkd2 mutations. It is concluded that a functional and molecular interaction exists between FPC and PC2 in vivo.

J Am Soc Nephrol 19: 455–468, 2008. doi: 10.1681/ASN.2007070770

Autosomal dominant polycystic kidney disease (AD- ADPKD affects 1 in every 500 to 1000 people and PKD) and autosomal recessive polycystic kidney dis- results from mutations in either of at least two ease (ARPKD) are common human genetic disorders causal genes, PKD1 and PKD2, leading to nearly that are characterized by numerous, expanding fluid- filled cysts in both kidneys and other duct/tubule-con- Received July 16, 2007. Accepted September 4, 2007. taining organs.1,2 ADPKD is inherited as a dominant trait, occurring relatively late in life, and is character- Published online ahead of print. Publication date available at www.jasn.org. ized by focal outpouchings of spherical cysts in the renal tubules. In contrast, ARPKD is inherited as a I.K. and Y.F. contributed equally to this work. recessive trait, usually presents during perinatal life, Correspondence: Dr. Guanqing Wu, Division of Genetic Medi- and is characterized by numerous spindle-shaped re- cine, Department of Medicine and Cell and Developmental Biol- ogy, Vanderbilt University, 539 LH, 2215 Garland Avenue, Nash- nal cysts. Because of these differences, ADPKD and ville, TN 37232. Phone: 615-936-1761; Fax: 615-936-2661; ARPKD are usually considered two distinct diseases in E-mail: [email protected] 3 clinical practice. Copyright © 2008 by the American Society of Nephrology

J Am Soc Nephrol 19: 455–468, 2008 ISSN : 1046-6673/1903-455 455 BASIC RESEARCH www.jasn.org

Figure 2. Survival analyses of Pkhd1 mutant mice. (A) Genotype and survival rate of Pkhd1 mutant mice from embryo to adult- ϩ Ϫ hood. (B) Kaplan-Meier survival curves for WT, Pkhd1 / , and Ϫ Ϫ Pkhd1 / mice are shown. The mice in these cohorts were ob- Ϫ Ϫ served for Ͼ1 yr. Survival of the Pkhd1 / mice differed signifi- ϩ Ϫ cantly from that of the WT and Pkhd1 / cohorts (P Ͻ 0.001).

identical clinical manifestations. Mutations in PKD2, which maps to chromosome 4q21–23, are responsible for approximately 15% of familial ADPKD cases.4,5 PKD2 has an approximately 5.4-kb transcript and encodes the 968–amino acid gene product polycystin-2 (PC2). PC2 is predicted to be an integral membrane protein with six putative transmembrane domains and intracellular NH2- and COOH-termini.4 PC2 has been reported to be a receptor-operated, nonselective cation channel; it is also referred to as TRPP2, because it is con- Figure 1. Pkhd1 gene-targeting construct and molecular analysis sidered a member of the trp superfamily.6,7 Conversely, muta- of the specific targeting event at the Pkhd1 locus. (A) Schematic tions in PKD1 are responsible for approximately 85% of representation of the gene-targeting strategy. (I) Partial genomic ⌬ ADPKD cases. PKD1 lies in an approximately 53-kb region on map showing exons 15 to 21 of Pkhd1. (II) Pkhd1e15GFP 16 target- chromosome 16p13.3 and yields a 14-kb transcript that en- ing vector in which exon 16 was deleted and exon 15 was dis- 8 rupted. (III) Partial map of the mutant Pkhd1 allele. Nh, NheI; S, codes a 4303–amino acid integral membrane protein. The SphI; P, PstI; B, BamHI; K, KpnI. (B) Tail biopsy DNA from mice gene product of PKD1, polycystin-1 (PC1), was reported to with the germline targeted mutation in Pkhd1 were digested with interact with PC2 and regulate Pkd2-channel activity.9–12 PstI and hybridized with probe 1 from outside the targeted region (AIII). The expected 3.3-kb WT band was observed in WT and Pkhd1 mutant alleles contained an in-frame GFP reporter gene, Pkhd1 heterozygous mice. A mutant 2.2-kb band was seen in we used an anti-GFP antibody to detect GFP expression. GFP Pkhd1 heterozygous and homozygous mice. (C) Tail biopsy DNA immunoreactivity was detected at the expected size of 130 kD in Ϫ Ϫ ϩ Ϫ were digested with BamHI and were hybridized with the Pkhd1 the adult kidneys of Pkhd1 / and Pkhd1 / mice. (F) Western exon 16 probe. The radioactive signal was not observed in Pkhd1 blot detection of FPC. Antibodies hAR-Cm3C10 and hAR- homozygous mice. (D) A 3.8-kb cDNA fragment containing exons Nm3G12, which recognize the COOH- and NH2-portions of FPC, 37 to 43 of Pkhd1 served as a Northern probe to detect Pkhd1 in respectively, were used to detect immunoreactivity to FPC in the Ϫ Ϫ ϩ Ϫ Ϫ Ϫ the total RNA of adult kidneys. Pkhd1 / mice showed an ab- adult kidneys of WT, Pkhd1 / , and Pkhd1 / mice. FPC expres- Ϫ Ϫ sence of Pkhd1 mRNA (D, top). The 28S rRNA band images sion was significantly reduced in Pkhd1 / kidneys compared ϩ Ϫ provided a total RNA loading control (D, bottom). (E) Because the with WT and Pkhd1 / .

456 Journal of the American Society of Nephrology J Am Soc Nephrol 19: 455–468, 2008 www.jasn.org BASIC RESEARCH

Figure 3. Hepatorenal cysts and tubular ectasia Ϫ Ϫ in Pkhd1 / mice. (A) A kidney tissue section shows normal glomeruli and a normal tubuloin- terstitial compartment in a 2-wk-old WT mouse. (B) Patchy dilation of proximal tubules and focal dilation of papillary collecting ducts were seen in Ϫ Ϫ a Pkhd1 / littermate. (C) There was a marked increase in the degree of tubular dilation with flattening of the tubular epithelial cells (arrow) in Ϫ Ϫ the kidney of a 4-wk-old Pkhd1 / mouse. (D) At 2 mo, there was a further increase in the severity of tubular dilation, in both the cortical proximal tubules and the medullary collecting duct tubules. Moreover, there was an increased expansion of Bowman’s space (arrow), an increase in mesangial cellularity, and a reduction of glomerular capillary loop formation. (E) At 4 mo, in the whole- Ϫ Ϫ mount Pkhd1 / kidney, there was a dramatic increase in tubular dilation, involving Ͼ80% of the cortex and medulla. (F) A high-power view of E at the cortical region of the kidney shows persistent expansion of Bowman’s space with a mild decrease in glomerular size and segmental mesangial hypercellularity. The proximal tubules were dilated. (G) The medullary sections of a higher-power view of E also showed occasional small cysts lined by a single layer of epithelium. Red blood cells were seen within the tubular lumen. (H) Liver section showed normal histology of the 2-wk-old WT mouse. (I) Liver section of the homozygous littermates showed dilation of the biliary ducts. (J) Tran- sillumination of an affected liver in the 2-mo-old Ϫ Ϫ Pkhd1 / mice showed cyst formation (left arrow) and dilated ducts (right arrow). (K) There were fo- cally enlarged cysts with areas of hemorrhage and necrosis (arrow) within the liver parenchyma in the same mouse. Bar ϭ 20 ␮minA,B,H,andI;10␮m in C, D, F, G, and K; and 1 mm in E and J.

ARPKD is one of the common hereditary renal cystic dis- tion,15 intracranial aneurysms,20,21 and adrenal insufficiency22 eases in infants and children.13 The estimated incidence of can be seen in patients with ARPKD. ARPKD is approximately 1 in 20,000 live births.14 The clinical ARPKD is caused by mutations in PKHD1. This gene con- characteristics of ARPKD include ectasia of the renal collecting sists of at least 86 exons spanning 470 kb on chromosome 6p12 ducts and hepatic biliary ducts with associated renal and he- and produces a 16-kb transcript. The longest open reading patic fibrosis.15 Approximately 50% of patients with ARPKD frame is predicted to include 66 exons and to encode the 4074– present with their disease as neonates16 and are born with two amino acid membrane-associated receptor-like protein fibro- very large kidneys with 60 to 90% of the renal tubules being cystin/polyductin (FPC).23–26 It was shown that FPC is associ- ectatic.15 These neonates suffer a 30% mortality rate as a result of ated with the basal bodies/primary cilia of epithelial cells27–30 respiratory and/or renal dysfunction.17 Patients who have and co-localizes with PC2 within the cell.31 These observations ARPKD and survive their first year of life have a more optimistic suggest the possibility that FPC and PC2 may function in a prognosis: 75% reach age 5, and only half develop ESRD.16–18 common molecular pathway in vivo. In rare cases, patients with ARPKD survive beyond age 60.19 To investigate a potential functional relationship between The most common complications of ARPKD include hy- FPC and PC2, we generated and characterized a mouse model pertension (60 to 100%), portal hypertension owing to severe with a gene-targeted mutation in Pkhd1. We discovered that hepatic fibrosis or Caroli disease (30 to 75%), and chronic lung lack of FPC was associated with the downregulation of PC2 disease (approximately 11%).17 In addition, growth retarda- expression in vivo. Renal cyst formation is more severe in mice

J Am Soc Nephrol 19: 455–468, 2008 FPC Regulates PC2 Expression and Function 457 BASIC RESEARCH www.jasn.org

Figure 4. Abnormal phenotypes were seen in Ϫ Ϫ the extrahepatorenal organs of Pkhd1 / mice. (A) Pancreatic sections from a 6-mo-old WT mouse showed normal histology. (B) In the same homozygous littermates, there was dilation of both the small and large pancreatic ducts (upper arrow) with a marked increase in interstitial fibrosis (lower arrow). (C) Normal brain section from a 2-mo-old WT mouse. (D) Brain tissue sections showed multiple, diffuse, small- to medium-sized vacuoles in a homozygous littermate (arrow). (E) The stomach mucosa of a 2-mo-old homozygous mouse showed an ulcerated lesion (arrow) that was raised with beaded borders. (F) The colon of the same mouse showed small areas of subse- rosal hemorrhage. (G) The mucosal aspect of the colon from a 2-mo-old homozygous mouse showed a small, ulcerated lesion with hemorrhage (arrow). (H) The microscopic colon section of the sample in G revealed a submucosal lymphocytic infiltrate with focal necrosis and erosion of the overlying epithelium. (I) Gross view of the small intestinal mucosa with mild superficial hemorrhage (arrow). Histologic sections of the sample in I showed submucosal edema, lymphocytic infiltrate, hemorrhage, and erosion of the overlying mucosa. Bar ϭ 25 ␮m in A through D; 5 ␮minH;1mMinE through G and I; and 10 ␮m in I inset.

Ϫ Ϫ that are transmutant in Pkhd1 and Pkd2 (Pkd2tm2Som)32,33 than Pkhd1 / mouse kidneys, suggesting that the homozygous in mice that are heterozygous for the Pkd2 mutation alone. In mice lacked full-length, functional FPC. addition, we found that Pkd2-channel activities were dysregulated in primary cultures of renal epithelial cells derived from Pkhd1-Deficient Mice Exhibit the Phenotypic Pkhd1 mutant mice. The in vivo observations demonstrated Characteristics of Human ARPKD that FPC and PC2 functionally interacted and resided in a To determine whether the Pkhd1-deficient mice would be a common molecular pathway. suitable model for ARPKD, we intercrossed heterozygous Pkhd1 mice to produce homozygous (Pkhd1Ϫ/Ϫ) progeny. Ϫ/Ϫ RESULTS Pkhd1 mice were born at a frequency of 18%, lower than the expected Mendelian ratio, indicating a loss of approxi- Ϫ Ϫ Generation of Pkhd1-Deficient Mutant Mice mately 5 to 10% of the Pkhd1 / mice during the embryonic or Because the Northern probe containing exons 15 to 16 of perinatal period (Figure 2A). This suggests that the Pkhd1Ϫ/Ϫ PKHD1 displayed highly mRNA expression signal in the kid- genotype may lead to embryonic lethality. Mice that reached ney tissue, we designed a targeting construct that not only dis- adulthood were used in a cohort study. Kaplan-Meier analysis rupted exon 15 but also deleted exon 16, resulting in a mutant showed that only 25% of Pkhd1Ϫ/Ϫ mice survived beyond 12 allele designated Pkhd1e15GFP⌬16 (Pkhd1Ϫ) (Figure 1,A mo compared with 60% of Pkhd1ϩ/Ϫ mice (P Ͻ 0.001) and through D). Because our gene-targeting construct included an 90% of WT mice (P Ͻ 0.001), suggesting that some adult in-frame green fluorescence protein (GFP) reporter gene with Pkhd1Ϫ/Ϫ mice died as a result of their disease phenotypes a predicted fusion protein of 130 kD (Figure 1A), anti-GFP (Figure 2B). Although there was a trend toward increased mor- immunoprecipitation assays showed that only the mutant tality in Pkhd1ϩ/Ϫ mice compared with WT mice, this did not mice expressed the GFP reporter protein (Figure 1E). In addi- reach statistical significance (P ϭ 0.075). tion, with the use of a panel of previously generated anti-FPC Pkhd1Ϫ/Ϫ mice that escaped embryonic lethality and sur- antibodies,31 Western blot analyses were able to detect the ex- vived into adulthood exhibited mild to severe tubular dilation pected immunoreactive bands at Ͼ400 kD in the wild-type or cyst formation in the kidney and liver accompanied by fi- (WT) and heterozygous mouse kidneys (Figure 1F). These im- brosis and necrosis (Figure 3). The severity of cysts and the age munoreactive bands were not detected in protein lysate from at the onset of disease varied among individual Pkhd1Ϫ/Ϫ mice.

458 Journal of the American Society of Nephrology J Am Soc Nephrol 19: 455–468, 2008 www.jasn.org BASIC RESEARCH

Figure 5. GFP expression in Pkhd1 mutant Ϫ Ϫ mice. (A) E15.5 Pkhd1 / kidney, liver, adrenal gland, and gastrointestinal tract stained with IHC using an anti-GFP polyclonal antibody. Positive signals were seen in the cortical adrenal cells (left arrow), periportal liver cells (right arrow), gastrointestinal tract (white box), and weakly positive staining was seen in the renal tubules (black box). (B) Higher magnification of the black box in A shows positive staining of the renal epithelia (arrow). (C) Higher magnification of the white box in A shows strong positive staining in the mucosal cells of the colon (arrow). (D) The differential interface contrast view showed GFP-positive staining (arrows) in the renal tubule epithelia ϩ Ϫ Ϫ Ϫ of 4-mo-old Pkhd1 / (D, top) and Pkhd1 / (D, bottom) littermates. (E) Positive GFP staining was detected in an alveolar bronchiole of the lung in a 2-mo-old homozygous mouse (arrow). (F) Positive GFP was also seen in the tracheal epithelium of the same mouse (arrow). (G) GFP-positive staining (red) appeared in the ependymal cells lining the ventricles of the brain in a 2-mo-old homozygous mouse (arrow). Yo-pro (green) was used to stain the nuclei of cells. (H) GFP-positive staining (red) was observed in the diseased liver of a 2-mo-old homozygous mouse (arrows). (I) Cytokeratin 7–positive staining (green), a marker for epithelial cells, outlined the biliary epithelial structures (arrows). (J) The merged confocal image showed that GFP co-localized with cytokeratin 7. Bar ϭ 30 ␮minA;15 ␮m in B and C; and 10 ␮m in D through J.

Approximately 10% of the mice had early-onset, microscopic of the cilia was shown to induce cyst formation in the kid- malformations in the renal tubules (Figure 3, B through G), neys34,35; therefore, we began to determine whether the lack of whereas others (approximately 60%) survived beyond 1 yr and FPC also disrupts ciliogenesis in Pkhd1-deficient mice. We had late-onset cystic phenotypes. The extent of cystic change in used IF with an anti-acetylated ␣-tubulin antibody to examine the liver of Pkhd1Ϫ/Ϫ mice was generally more severe than that the number and morphology of renal primary cilia in 6-mo- observed in the kidneys (Figure 3, H through K). These obser- old littermates. Compared with WT mice, there were far fewer vations indicate that the targeted mutation in Pkhd1 induces primary cilia in the renal tubular epithelia of Pkhd1Ϫ/Ϫ mice cystogenesis in the kidneys and livers of the mice. Cystic or (Figure 6, A versus B), indicating that lack of FPC might reduce dilated-duct phenotypes were also seen in the pancreas (Figure ciliogenesis in the kidneys. The ciliary defects seem to be most 4, A versus B) and brain (Figure 4, C versus D) of Pkhd1Ϫ/Ϫ severe in the cortical proximal tubules of the Pkhd1Ϫ/Ϫ kid- mice. In the gastrointestinal tract, hemorrhagic, and ulcer-like neys. In addition, confocal images also displayed similar ciliary lesions were observed (Figure 4, E through I). changes in corresponding regions of littermate kidneys (Figure Because the Pkhd1 mutant allele carries an in-frame GFP 6, C versus D). Scanning electron microscopy confirmed these reporter gene in exon 15, we used immunohistochemistry results in 3-mo-old littermates with or without the targeted (IHC) and immunofluorescence (IF) staining with anti-GFP Pkhd1 mutation (Figure 6, E versus F). For further validation of antibodies to detect GFP distribution in the organs of homozy- these findings, primary culture of renal epithelial cells derived gous mutant mice. GFP-positive signals, defining the expres- from 2-mo-old Pkhd1Ϫ/Ϫ and WT littermates was used to deter- sion pattern of Pkhd1, were observed in the tubular epithelia of mine whether the ciliary malformation could be observed in vitro. the kidneys (Figure 5, A, B, and D), gastrointestinal tract (Fig- Compared with the cultured WT cells, a shortened ciliary struc- ure 5, A and C), bronchioles/trachea (Figure 5, E and F), ture and decreased ciliary staining were seen in the Pkhd1Ϫ/Ϫ cells ependymal cells lining the ventricles of the brain (Figure 5G), (Figure 6, G versus H). The cilia stained in approximately 94% of and hepatobiliary epithelial cells (Figure 5, H through J). WT cells and in fewer than 34% of Pkhd1Ϫ/Ϫ cells (P Ͻ 0.001; Figure 6I). The mean length of primary cilia was 10 ␮m in cul- Ϫ Ϫ Lack of FPC Exhibits Aberrant Ciliogenesis in the tured WT cells and was Ͻ5 ␮minPkhd1 / littermate cells (P Ͻ Renal Epithelial Cells 0.001; Figure 6J). That short or absent cilia were observed in FPC was demonstrated to localize to the primary cilium and/or Pkhd1Ϫ/Ϫ cells but not in WT cells suggests that lack of FPC causes basal bodies of renal tubular epithelia,27–31 and malformation defects in ciliogenesis in renal epithelial cells in vivo.

J Am Soc Nephrol 19: 455–468, 2008 FPC Regulates PC2 Expression and Function 459 BASIC RESEARCH www.jasn.org

Ϫ Ϫ Figure 6. Lack of FPC induces aberrant ciliogenesis in the Pkhd1 / kidneys. (A) A common ciliary marker, anti-acetylated ␣-tubulin Ϫ Ϫ antibody, was used for IF staining of kidney sections from 6-mo-old WT and Pkhd1 / littermates. Ciliary structures (arrows) were abundantly observed in the WT kidneys. (B) Decreased ciliary staining (arrows) was seen in the corresponding cortical region of Ϫ Ϫ Pkhd1 / mouse kidneys. (C) The confocal images also showed normal ciliary structures in the 4-mo-old WT kidneys (arrows). Lotus Ϫ Ϫ Tetragonolobus Lectin (green) was used to stain renal proximal tubules. (D) The ciliary structures in the Pkhd1 / littermates were reduced in number and shorter (arrow) than WT controls. (E) Scanning electron microscopy showed the normal primary cilia (arrows) of Ϫ Ϫ a 3-mo-old WT kidney. (F) The ciliary structures in the Pkhd1 / kidney are shorter than those in the WT littermates, in the similar regions Ϫ Ϫ of the kidney. (G) Primary cultures of renal epithelia derived from the 2-mo-old WT and Pkhd1 / kidneys were stained with an anti-acetylated ␣-tubulin antibody. The confocal images showed normal ciliary structures (arrows) in both top (G, top) and lateral views Ϫ Ϫ (G, bottom). Blue To-pro was used to stain nuclei. (H) In the primary cultured epithelial cells from the Pkhd1 / littermate, confocal images showed ciliary structures that were shorter and fewer in number than those in control cells (arrows). The confocal top and lateral views were composed of multiple sections (approximately 0.5 ␮m thick and up to 16 layers) that were projected onto one plane to present the ciliary staining patterns. (I) One hundred individual primary cultured cells from five random high-power fields (ϫ1000) were numbered; the cell number and positive cilium-staining rates are shown in I. In WT cells, 94% of cells stained positive for cilia, compared Ϫ Ϫ with 34% of Pkhd1 / littermate cells (P Ͻ 0.001). (J) The length of 50 individual primary cilia of cultured cells from three random high-power fields was measured using lateral views of the confocal images; the average length of the primary cilia was calculated and Ϫ Ϫ showed in J. The primary cilium length is approximately 10 ␮minWTandϽ5 ␮minPkhd1 / littermate cells (P Ͻ 0.005). Bar ϭ 10 ␮m in A through D; 2 ␮m in E and F; and 5 ␮m in G and H.

Trans-Mutant Mice in Pkhd1 and Pkd2 Accelerate Because spherical renal cysts (at least three times the normal Renal Cyst Disease Progression diameter of the proximal tubule) are characteristic of ADPKD Several in vitro studies from our group and others have shown and massive, dilated, spindle-shaped renal tubules are more that FPC and PC2 co-localize to the basal bodies/primary cilia typical of the ARPKD phenotype, we used sphere-shape cysts Ϫ of renal epithelial cells and are able to form a molecular com- to represent ADPKD-like renal cysts.1–3 In mice with Pkhd1 / ϩ Ϫ plex and function in the same signaling pathway.31,36 For fur- Ϫ/Pkd2 / alleles, spherical cysts were distributed at the med- ther validation of this finding in vivo, the phenotypic effects of ullary and cortical regions on a background of Pkhd1Ϫ/Ϫ-spe- transheterozygosity for Pkhd1 and Pkd2 were examined. We cific dilated tubules (Figure 7A). By statistical analysis, Pkhd1Ϫ/ ϩ Ϫ ϩ Ϫ ϩ Ϫ intercrossed Pkhd1 / and Pkd2 / mutant mice to produce Ϫ/Pkd2 / mice had significantly greater numbers of spherical cohorts of age-matched littermates. The four genotypes of in- cysts than other genotypes (P Ͻ 0.05; Figure 7B). Western blot terest (Pkhd1Ϫ/Ϫ, Pkhd1Ϫ/Ϫ/Pkd2ϩ/Ϫ, and WT) were obtained analyses of lysates from Pkd2ϩ/Ϫ and Pkhd1Ϫ/Ϫ/Pkd2ϩ/Ϫ litter- from live-born progeny. Cohorts were killed at 1 mo. Gross mates demonstrated that loss of FPC in the adult mouse kidney inspection did not reveal cysts on the surface of the kidneys in reduced PC2 expression in vivo (Figure 7C). These studies indi- any of the genotypes. cate that lack of FPC exacerbates the severity of ADPKD.

460 Journal of the American Society of Nephrology J Am Soc Nephrol 19: 455–468, 2008 www.jasn.org BASIC RESEARCH

termate kidneys (data not shown), suggesting that lack of FPC may affect only the synthesis and/or stability of PC2 in vivo. To investigate whether FPC and PC2 physically interact, we used tissue lysates from E13.5 WT, Pkhd1ϩ/Ϫ, and Pkhd1Ϫ/Ϫ littermates to perform a co-immunoprecipitation assay (co- IP) with antibodies against FPC and PC2. We found that FPC immunoprecipitated with PC2 in tissue from WT and Pkhd1ϩ/Ϫ littermates but not from the negative control Pkhd1Ϫ/Ϫ littermates (Figure 8C). In addition, we used tissue lysates from an E13.5 Pkd2-mutant set in another co-IP assay. Similar to the previous result, FPC immunoprecipitated with PC2 in tissue from WT and Pkd2ϩ/Ϫ littermates but not in tissue from Pkd2Ϫ/Ϫ littermates (Figure 8D). Immunoreactiv- ity was stronger in WT than in heterozygous littermates and was not detected in homozygous littermate controls, providing strong evidence that FPC may physically interact with PC2 in vivo. It is interesting that no reduction of FPC expression was seen in Pkd2Ϫ/Ϫ littermates by Western analysis, suggesting that lack of PC2 does not affect the level FPC in vivo (Figure 8E). We constructed serial Hemagglutinin (HA) and Flag-tagged Figure 7. The cystic phenotype of mice trans-mutant for Pkhd1 expression vectors that contain intracellular portions of FPC and and Pkd2. (A) In a representative hematoxylin- and eosin-stained PC2, respectively. Positive immunoreactivity was seen in co-IP section, spherical renal cysts (arrows; diameter Ͼ50 ␮m was con- assays between the COOH-terminal portion of FPC (FPC-C- Ϫ Ϫ sidered as renal cyst) were identified in a 1-mo-old Pkhd1 / / Flag) and the NH2-terminal portion of PC2 (PC2-N-HA; Figure ϩ Ϫ Pkd2 / double-mutant mouse. (B) Numbers of cysts are pre- 8F). Immunoreactivity was not observed between COOH-termi- sented as means Ϯ SD for four genotypes at the age of 1 mo (n ϭ nal portions of FPC (FPC-C-Flag) and PC2 (PC2-C-HA; data not number of animals in each group). The increase in spherical renal shown). This suggests that the interaction between FPC and PC2 Ϫ/Ϫ ϩ/Ϫ cysts in Pkhd1 /Pkd2 trans-mutant mice was significantly occurs via the intracellular COOH-terminal tail of FPC and NH2- higher than other genotypes (*P Ͻ 0.05). (C) Western blot of ϩ Ϫ terminal portion of PC2. For confirmation of this, HEK293 cells protein lysates from Pkd2 / 1-mo-old mouse kidney with or Ϫ Ϫ were transiently co-transfected with both an expression vector without the Pkhd1 / mutation was performed using the anti-PC2 antibody hPKD2-Cm1A11. A significant decrease in immunoreac- containing the full-length human PKD2 cDNA (PC2-Full) and Ϫ Ϫ tivity in lysates from the Pkhd1 / mutant mice indicates that lack the aforementioned FPC-C-Flag vector. With the use of an anti- of FPC reduces PC2 expression in vivo. Bar ϭ 30 ␮minA. PC2 antibody to immunoprecipitate and an anti-Flag antibody for detection by Western blot, a strong band was seen in lysate FPC and PC2 Interact In Vivo and Form a Molecular from the co-transfected cells. Weak immunoreactivity was ob- Complex in Renal Epithelia served in lysate from cells that were transfected only with FPC-C- Because a lack of FPC inhibited the expression of PC2 and Flag. These data suggest that FPC-C-Flag can be co-immunopre- induced severe cystic phenotypes in the Pkhd1Ϫ/Ϫ/Pkd2ϩ/Ϫ cipitated with either PC2 introduced exogenously or PC2 mouse model (Figure 7), we decided to analyze the molecular endogenous to HEK293 cells (Figure 8G). In addition, when ly- relationship between FPC and PC2. We used our mAb hPKD2- sates from FPC-C-Flag–transfected HEK293 cells were preincu- Cm1A11, which is directed against the intracellular COOH- bated with an antibody against the COOH-terminal tail of FPC terminus of PC2, to examine protein level in tissue lysates from (hAR-C2p), the detection of PC2-N-HA and FPC-C-Flag co-IP WT, Pkhd1ϩ/Ϫ, and Pkhd1Ϫ/Ϫ littermates at embryonic day was blocked (Figure 8H). This result provides further evidence of 13.5 (E13.5). Compared with WT embryos, PC2 expression a specific interaction between the C-terminal tail of FPC and the was decreased in the Pkhd1Ϫ/Ϫ embryos, suggesting that lack of N-terminal portion of PC2. FPC downregulates PC2 expression in vivo (Figure 8A). In addition, IHC staining with the anti-PC2 polyclonal antibody Lack of Pkhd1 Disrupts Pkd2-Induced Cation Channel hPKD2-Cp was used to examine expression level of PC2 in the Activities Pkhd1Ϫ/Ϫ 1-mo-old kidneys. In the renal cortex, significantly To study further whether FPC and PC2 functionally interact, less staining was observed in Pkhd1Ϫ/Ϫ kidneys than in WT we used whole-cell patch clamp recordings to characterize littermates (Figure 8B), giving further evidence that lack of Pkd2-channel activities in Pkhd1-deficient cells. We first tested FPC disrupts normal PC2 expression in vivo; however, quan- whether Pkd2-induced whole-cell current could be detected titative PCR did not reveal a significant difference in Pkd2 ex- under published conditions37,38 and then investigated whether pression among Pkhd1Ϫ/Ϫ, Pkhd1ϩ/Ϫ, and WT 1-mo-old lit- the recorded channel-like current is Pkd2-specific. First, pri-

J Am Soc Nephrol 19: 455–468, 2008 FPC Regulates PC2 Expression and Function 461 BASIC RESEARCH www.jasn.org

Figure 8. Molecular relationship between FPC and PC2. (A) Using the anti-PC2 mAb hPKD2-Cm1A11, Western blot of duplicate ϩ Ϫ Ϫ Ϫ Ϫ Ϫ protein lysates from WT, Pkhd1 / , and Pkhd1 / E13.5 littermates showed a significant downregulation of PC2 in Pkhd1 / embryos, indicating that lack of FPC inhibits PC2 expression in vivo. An anti–␤-actin antibody was used for a protein-loading control. (B) In comparison with the WT littermate (left), IHC staining with the anti-PC2 polyclonal antibody hPKD2-Cp showed a significant decrease Ϫ Ϫ ϩ Ϫ in PC2 expression in the cortical region of the 1-mo-old Pkhd1 / kidney (right). (C) Duplicate lysates from E13.5 WT, Pkhd1 / , and Ϫ Ϫ Pkhd1 / littermates were used to perform a co-IP Western using the anti-FPC antibody hAR-Nm3G12 to IP and the anti-PC2 antibody hPKD2-Cm1A11 to detect PC2 expression. Positive immunoreactivity was seen in the WT embryo, and progressively reduced ϩ Ϫ Ϫ Ϫ immunoreactivities were seen in the Pkhd1 / and Pkhd1 / littermates, suggesting that FPC binds to PC2 in vivo. (D) Lysates from ϩ Ϫ Ϫ Ϫ E13.5 WT, Pkd2 / , and Pkd2 / littermates were used to perform a co-IP Western using the anti-PC2 antibody to IP and the anti-FPC ϩ Ϫ antibody to detect FPC expression. Positive immunoreactivity was seen in the WT and Pkd2 / littermates, but no immunoreactivity was Ϫ Ϫ observed in the Pkd2 / littermate, providing further evidence that FPC physically interacts with PC2 in vivo. (E) There was no change Ϫ Ϫ in FPC expression in Western blot analysis among the WT and Pkd2 / littermates, indicating that the downregulation of PC2 does not affect FPC expression. (F) HA- and Flag-tagged expression vectors, in which the COOH-terminus of FPC (FPC-C-Flag) and the NH2-terminus of PC2 (PC2-N-HA) were constructed in-frame, were transiently co-transfected into HEK293 cells. Using an anti-Flag antibody to IP and an anti-HA antibody to detect the NH2-terminus of PC2, positive immunoreactivity was seen only in the co-transfected sample, indicating that the COOH-terminus of FPC physically interacts with the NH2-terminus of PC2 in vitro. (G) The same FPC-C-Flag expression vector was transiently co-transfected into HEK293 cells with an expression vector containing the human full-length PKD2 cDNA (PC2-Full). The anti-PC2 antibody hPKD2-Cm1A11 was used for IP and an anti-Flag antibody was used to detect the COOH-terminus of FPC. Strong positive immunoreactivity was seen only in the co-transfected sample, and weak immunoreactivity was detected in the FPC-C-Flag single-transfected sample, indicating that either exogenously transfected or endogenously expressed PC2 immunoprecipitates with FPC-C-Flag construct. This further confirms that the COOH-terminus of FPC physically interacts with the NH2-terminus of PC2. (H) Using the hAR-C2p antibody against the COOH-terminus of FPC to preincubate with FPC-C-Flag single- transfected protein lysates, positive immunoreactivity was seen only in the nonpreincubated co-IP sample, whereas the immunoreactivity was missing in the preincubated co-IP sample. Bar ϭ 30 ␮minB. mary cultured fibroblasts derived from E13.5 WT and Pkd2Ϫ/Ϫ by missing Pkd2 expression (P Ͻ 0.005; Figure 9A). Next, we embryos were tested to determine whether whole-cell current used the same approach to test the Pkhd1-silenced inner med- densities were dysregulated by the lack of Pkd2. The recorded ullary collecting duct (IMCD) cells (IMCDshRNA3e23) and WT Ϫ Ϫ current from the Pkd2 / fibroblasts was significantly lower control cells (IMCDsh15) that we had generated previously.39 It than that from the WT fibroblasts at all voltages tested, indi- is interesting that Pkhd1-silenced IMCD cells exhibited a sim- cating that reduction of functional channel activities is caused ilar channel reduction, indicating that inhibition of FPC also

462 Journal of the American Society of Nephrology J Am Soc Nephrol 19: 455–468, 2008 www.jasn.org BASIC RESEARCH

Figure 9. Analysis of whole-cell current recordings in Pkhd1-deficient cells. (A) Average current density (pA/pF) and voltage relation (I-V Ϫ Ϫ curve) of primary cultured fibroblast cells from E13.5 embryos show a statistically significant difference between WT and Pkd2 / littermates, suggesting that the cation channel activity is induced by lack of PC2 (n ϭ 10; *P Ͻ 0.005). (B) A Pkhd1-knockdown stable cell line IMCDshRNA3e23 and its WT–controlled cell line IMCDsh15 were used to perform the whole-cell current recording assays under the same conditions. Currents were elicited by stepping from a holding potential of 0 mV to various test potentials (Ba). The current densities between IMCDshRNA3e23 (n ϭ 12) and IMCDsh15 (n ϭ 7) cells exhibited a statistically significant difference (Bb versus Bc; *P Ͻ 0.001). The control IMCDsh15 cells displayed a steep inwardly rectifying current, which was significantly reduced by the intracellular introduction of the anti-PC2 C-terminal antibody hPKD2-Cp via pipette solution at a dilution of 1:200 (n ϭ 8; Bb versus Be; #P Ͻ 0.02), suggesting that the current changes are PC2 specific. (C) Primary cultured renal epithelial cells derived from 2-mo-old WT (WT) and Ϫ Ϫ Pkhd1 / littermates were used to perform the same whole-cell current recording assays. The control WT cells (n ϭ 15) displayed a Ϫ Ϫ steep inwardly rectifying current, which was substantially reduced in Pkhd1 / cells (n ϭ 7). This indicates that the whole-cell current Ϫ Ϫ density in cells from the Pkhd1 / mice is significantly lower than in those from the WT littermates (*P Ͻ 0.003). (D) hPKD2-Cp and a negative control anti-actin antibody were each added intracellularly into separate WT cells via pipette solution. Representative current traces were not clearly reduced in the cells with anti-actin antibody (n ϭ 6), but a substantially smaller current amplitude was seen in Ϫ Ϫ the cells with hPKD2-Cp antibody (n ϭ 8; #P Ͻ 0.01), further suggesting that the current alteration between WT and Pkhd2 / cells in Ϫ Ϫ CisPkd2 specific. (E) For unequivocal validation that the inhibited current density seen in the Pkhd1 / cells is due to lack of FPC expression, a human full-length PKHD1 cDNA was in-frame constructed into a GFP-tagged expression vector and was transiently Ϫ Ϫ transfected into the Pkhd1 / cells. GFP-positive cells were chosen to perform the whole-cell current recording (n ϭ 4), and current Ϫ Ϫ density-voltage plots showed that the re-expression of FPC rescues the inhibited current density seen in the Pkhd1 / cells Ϫ Ϫ Ϫ Ϫ Ϫ Ϫ (Pkhd1 / -C versus Pkhd1 / ;#P Ͻ 0.01). The dashed line represents the mean I-V curve from the Pkhd1 / cells in C, and the dotted line shows the means I-V curve from the WT cells in C. Currents were measured 220 ms after stepping to the test potential. Data are means Ϯ SEM from the tested independent cells. Statistical difference at *ϩ80 and #ϩ100 mV between the tested cell groups. The data ϭ Ϫ (V50 Ϫ V)/k) ϩ were fitted with standard Boltzmann equation (I Imax/1 e C). reduces Pkd2-channel activities (P Ͻ 0.001; Figure 9, Bb versus antibody (P Ͻ 0.02; Figure 9, Bb versus Be), suggesting that the Bc). For verification that the functional channel changes were current change is Pkd2 specific. To validate our findings fur- induced by missing Pkd2, the anti-PC2 polyclonal antibody ther, we produced primary cultures of renal epithelia from hPKD2-Cp, which recognizes the intracellular COOH-termi- 2-mo-old WT and Pkhd1Ϫ/Ϫ kidneys. Whole-cell current den- nal region of PC2, was transferred into WT control IMCDsh15 sities were significantly decreased in Pkhd1Ϫ/Ϫ cells compared cells via pipette solution. The current density was significantly with WT cells (P Ͻ 0.003; Figure 9C), indicating that lack of inhibited in IMCDsh15 cells that were treated with hPKD2-Cp FPC also inhibits Pkd2-specific channel activities. For clarifi-

J Am Soc Nephrol 19: 455–468, 2008 FPC Regulates PC2 Expression and Function 463 BASIC RESEARCH www.jasn.org cation that the downregulated current in Pkhd1Ϫ/Ϫ primary our findings combined with theirs draw the conclusion that cultured cells was due to the Pkd2-specific channel, the whole- ADPKD and ARPKD may reside in the same pathogenic path- cell current of WT primary cultured cells was recorded in the way. presence of either the anti-PC2 antibody hPKD2-Cp or an an- In this study, the Pkhd1Ϫ/Ϫ genotype seemed to cause em- ti-actin antibody. As representative current records and I-V bryonic lethality; however, a phenotypic analysis of these curves in Figure 9D show, the anti-PC2 antibody reduced the Pkhd1Ϫ/Ϫ embryos did not show any significant cardiac defects current density, but the anti-actin antibody did not, confirm- and only mild edema, which was prevalent in the Pkd2 mutant ing that the whole-cell current was conducted through Pkd2 mice. Because Pkd2Ϫ/Ϫ mice are embryonically lethal, the find- channels. ing that a lack of FPC promotes a significant downregulation of To obtain unequivocal evidence that the whole-cell current PC2 expression raises the possibility that partial embryonic changes were induced by downregulation of FPC, we further lethality in Pkhd1Ϫ/Ϫ mice may be caused by this reduction of investigated whether re-expression of FPC could rescue the PC2. reduced channel activities seen in Pkhd1Ϫ/Ϫ cells (Figure 9C). We previously reported that FPC and PC2 co-localize to the We transiently transfected primary cultured Pkhd1Ϫ/Ϫ cells same subcellular organelles, the basal bodies/primary cilia of with a mammalian expression vector, pcDNA3.1/Hygro, con- renal epithelial cells,31 suggesting that these two cystoproteins taining the full-length ORF cDNA of human PKHD1 with may form a molecular complex. Recent reports have shown GFP fused in-frame. GFP-positive Pkhd1Ϫ/Ϫ cells, named that the same chemical molecule (a vasopressin V2 receptor Pkhd1Ϫ/Ϫ-C, were chosen for the whole-cell current recording antagonist, OPC31260) can inhibit cyst progression in both a assay. As shown in Figure 9E, both the current density and I-V rat genetic model of ARPKD and a mouse genetic model of relation of Pkhd1Ϫ/Ϫ-C cells were used to compare the cur- ADPKD, suggesting that both causal gene products for AD- rents recorded from WT (dotted line) and Pkhd1Ϫ/Ϫ cells PKD and ARPKD may reside in a common molecular path- Ϫ Ϫ (dashed line; Figure 9C). Unlike Pkhd1 / cells, the current way.42 Recently, our company study demonstrated that FPC density of Pkhd1Ϫ/Ϫ-C cells was similar to WT cells (Figure and PC2 indirectly interact via their COOH-termini and that 9E), indicating that the current changes between WT and this is mediated by KIF3B, a motor subunit of the heterotrimer Ϫ Ϫ Pkhd1 / cells are due to downregulation of FPC. That inhi- kinesin-2.36 Pkd2-channel activities were significantly altered bition of Pkhd1 expression disrupts Pkd2-specific channel ac- when the FPC–PC2 complex was disrupted. This in vitro study tivities gives further evidence to demonstrate FPC and PC2 provides a molecular basis for a functional link between FPC function in the same molecular pathway. and PC2. Another recent report that supports a functional link suggests that FPC regulates mechanotransduced Ca2ϩ responses, which may be induced by PC2, in cultured Pkhd1- DISCUSSION knockdown cells.43 In this study, the evidence that the COOH- terminus of FPC directly interacts with the NH2-terminus of PC2 Although the gene responsible for ARPKD, PKHD1, has been suggests that FPC and PC2 are able physically to form a het- identified23–25 and its gene product, FPC, has been initially erodimeric complex in vivo. Lack of FPC downregulates Pkd2- characterized,27,28,30,31,39,40 the mechanisms by which PKHD1 channel activities in either the Pkhd1-knockout renal epithelial causes disease phenotypes remain largely unknown. To study cells in primary culture or the Pkhd1-knockdown IMCD cells that the disease mechanism and pathogenesis of ARPKD, we cre- we generated previously.39 Given that FPC and PC2 physically ated a mouse that allows manipulation of Pkhd1, an animal interact and that the lack of FPC downregulates PC2 expression in model that recapitulates the human ARPKD phenotype. vivo but PC2 does not downregulate FPC, we speculate that PC2 Through genetic and biochemical studies, we demonstrated may function immediately downstream of FPC. that the COOH-terminus of FPC physically interacts with the The disruption of ciliary formation in renal epithelia in- NH2-terminus of PC2 and that lack of FPC leads to downregu- duces cystogenesis in the kidneys.34,35 We recently reported lation of PC2 expression in vivo. Transmutant mice for Pkhd1 that downregulation of Pkhd1 significantly decreases ciliary and Pkd2 displayed a significantly more severe renal cystic phe- formation in cultured Pkhd1-silenced IMCD cells, suggesting notype than single-mutant mice, suggesting that Pkhd1 serves that lack of FPC might disrupt ciliogenesis in renal epithelial as a disease modifier for ADPKD. For functional validation of cells. This result is consistent with studies in which transient these findings, Pkd2-channel activities were examined both in small interference RNA–mediated inhibition of Pkhd1 in primary cultures of renal epithelia derived from Pkhd1 mutant cholangiocytes resulted in shortening and decreased formation mice and in Pkhd1-silenced IMCD cells.39 Pkd2-channel activ- of cilia,29 but spatial and environmental differences between in ities were significantly dysregulated in Pkhd1-deficient cells, vivo tissues and in vitro cell culture may lead to different results. indicating that FPC and PC2 reside in the same molecular Mouse models with a deletion of Pkhd1 exon 4044 and gene- pathway. During submission of our article, another group re- targeted mutations in Pkd2 that cause distinct liver and/or kid- ported that there are genetic interactions between Pkhd1 and ney cysts33,45,46 do not exhibit defects in ciliary structure in the Pkd1.41 Because the gene product of PKD1, PC1, was reported affected epithelial cells, suggesting that the failure of renal ep- to interact with PC2 and regulate Pkd2-channel activity,9–11 ithelia to assemble primary cilia may not be the only factor

464 Journal of the American Society of Nephrology J Am Soc Nephrol 19: 455–468, 2008 www.jasn.org BASIC RESEARCH leading to cyst formation in the kidneys. For example, a recent but not vice versa suggest that PC2, also known as TRPP2, study showed that disruption of the extracellular matrix pro- functions immediately downstream of FPC. In addition, be- tein laminin ␣5, which is a major component of the tubular cause inhibition of FPC expression reduces Pkd2-channel ac- and glomerular basement membranes, also produces cystic tivity, we conclude that a functional and molecular interaction kidneys in Lama5 mutant mice.47 exists between FPC and PC2 in vivo. Extracellular biochemical By carefully examining ciliogenesis in our Pkhd1 mutant and/or physical signals may activate the receptor-like protein kidneys, we found that primary cilia of the cortical tubular FPC, which then triggers TRPP2 channel to transmit signals epithelia were reduced in number and were shorter than con- that affect intracellular processes. Our in vivo study reveals an trols, documenting that disruption of FPC expression induces intriguing molecular relationship between FPC and PC2. malformation of the primary cilia in the kidneys in vivo. Re- cently, an elegant study demonstrated that disruption of FPC expression causes defective planar cell polarity in another CONCISE METHODS Pkhd1 genetic model, the pck rat. Spindle orientation in the renal epithelial cells of the pck rat is aberrant, suggesting that Mouse Strains cell polarity is disrupted during mitosis of renal epithelial The gene-targeted mouse model for Pkd2 (Pkd2tm2Som) was cells.48 Our previous in vitro study also demonstrated that renal previously generated by us.32,33 To produce mutant mice for epithelial IMCD cells with downregulated FPC exhibit aber- Pkhd1, we designed a targeting construct disrupting its 15th rant migratory polarity and lose the ability to drive collective coding exon (Figure 1). We found 620 embryonic stem cell cell migration, suggesting that the planar cell polarity also colonies resistant to G418, with one (W4A5) identified by PCR might be disrupted.39 Aberrant planar cell polarity seen in cells screening using a pair of outside-construct and cassette-based with Pkhd1 defects suggests that ciliary defects in our Pkhd1 primers. This cell line was further confirmed by Southern blot mutant mice might be caused by impeding orientally centriole analysis and injected into C57Bl/6 blastocysts at the gene tar- arrangement and disabling the establishment of epithelial po- geting and transgenic facility of University of Connecticut larity.39,49 Health Center. Because our Pkhd1 mutant mice bear an in-frame GFP reporter, immunostaining with anti-GFP antibodies provided an Southern and Northern Blotting and Quantitative PCR FPC expression profile in affected tissues. Positive GFP expres- Southern analysis was used to genotype Pkhd1 and Pkd2 mu- sion was detected in the apical domain of epithelial derivatives, tant mice with our published approaches.32,52 For Northern including the renal, hepatic, pulmonary, and gastrointestinal analysis, total RNA was isolated from embryos or kidneys us- epithelia as well as ependymal cells lining the ventricles of the ing Trizol reagent (Invitrogen, Carlsbad, CA) following the brain. These findings are consistent with our previous report31 manufacturer’s instructions. Probes were labeled using the and indicate that FPC may modulate the morphogenesis and RadPrime DNA-labeling system (Invitrogen) with ␣-32P- maintenance of tubular/ductal architectures in organs gener- dCTP (PerkinElmer, Waltham, MA) and were hybridized with ated from the primary duct system.50 Several recent studies total RNA blots (25 ␮g/lane). Images of 28-S rRNA bands in indicated that FPC is involved in notch-like processing and these same blots were used as a total RNA loading control. that its extremely large extracellular domain is released from Quantitative PCR was performed using the iCycler iQ Real- the primary cilia of renal epithelial cells.51 That GFP-positive Time PCR Detection System with iQ SYBR Green Supermix kit signals were detected at the microvilli of tubular epithelia in (Bio-Rad, Hercules, CA). Two pairs of primers were designed Ϫ Ϫ our Pkhd1 / mice agrees with the report that FPC may be from each cDNA sequence of Pkhd1 and Pkd2 (Table 1). released into the tubular/ductal lumen. In summary, we produced a mouse model for Pkhd1 that Antibodies recapitulates the phenotypic characteristics of human ARPKD. Polyclonal antibodies and mAb against FPC (including hAR- Using this model along with a Pkd2 mutant mouse, we were Np, hAR-C2p, hAR-Nm3G12, and hAR-C2m3C10) and anti- able to demonstrate the importance of FPC expression for nor- bodies against human PC2 (hPKD2-Cp and hPKD2-Cm1A11, mal PC2 function. The finding of aberrant ciliogenesis in the formerly named PKD2A11) were described in our previous kidneys of Pkhd1-deficient mice indicates that FPC disrupts studies.31,36 In addition, other polyclonal antibodies and mAb the process of ciliogenesis. That FPC physically interacts with were purchased: Anti-acetylated ␣-tubulin, anti–␥-tubulin, PC2 and that lack of FPC destabilizes PC2 expression in vivo anti–␤-actin, anti-Flag, and anti-HA mAb (Sigma, St. Louis,

Table 1. Primers for quantitative PCR Genes Primer Names Location Forward Primers Reverse Primers Pkhd1 Pkhd1-P1 Exons 18 to 19 5Ј-atgtctccagccaaccagttcc-3Ј 5Ј-gccttctaaaccttgctcaaatcc-3Ј Pkhd1-P2 Exon 65 5Ј-accttcgttgtcttgcc-3Ј 5Ј-tctggttttgcttttct-3Ј Pkd2 Pkd2-P1 Exons 3 to 4 5Ј-gacagagtcagtctttccatcgttc-3Ј 5Ј-acgcggcactcctagcag-3Ј Pkd2-P2 Exons 14 to 15 5Ј-tttctaagattgacgccgtg-3Ј 5Ј-gcttacaccatgacctgtttgc-3Ј

J Am Soc Nephrol 19: 455–468, 2008 FPC Regulates PC2 Expression and Function 465 BASIC RESEARCH www.jasn.org

MO); anti-GFP polyclonal antibodies (ab6556; Abcam, Cam- Statistical Analyses bridge, MA); fluorescein lotus tetragonolobus lectin (Vector The survival rate was determined by observing the cohorts of Laboratories, Burlingame, CA); anti–cytokeratin 7 polyclonal mice daily. We analyzed the survival using the Kaplan-Meier antibody (Santa Cruz Biotechnology, Santa Cruz, CA); and the function in the R Software. Graphical data are presented as nucleic acid dyes To-pro and Yo-pro (Molecular Probes, Eu- means Ϯ SD. Statistical analysis was performed where appro- gene, OR). priate using the t test or one-way ANOVA followed by Tukey multiple comparison test. Differences with P Ͻ 0.05 were con- Western Blotting and Immunoprecipitation sidered statistically significant. For Western analysis and immunoprecipitation, the detailed approaches were similar to our previous publications.31,36,39 For performance of co-IP and Western analysis, the entire in- ACKNOWLEDGMENTS tracellular termini of FPC and PC2 were constructed into Flag- tagged and HA-tagged pCMV expression vectors53 and were This work was supported by grants from the National Institutes of named FPC-C-HA (amino acids 3872 to 4074), PC2-N-HA Health (DK062373 and DK071090) to G.W. (amino acids 1 to 221), and PC2-C-Flag (amino acids 682 to We kindly thank Dr. Stefan Somlo for allowing us to use the PKD2 tm2Som 968). mutant model (Pkd2 ) that was generated in his laboratory. We also thank Dr. James P. Smith for excellent advice and suggestions; Dr. Caiying Guo for the embryonic stem cell work at University of Con- Histology, IF Staining, IHC, Confocal Microscopy, and necticut; Dr. Chun Li for statistical analysis; and Drs. Sae-youll Cho, scanning electron microscopy Shun-Wei Huang, Aijun Zuo, and Hong Wang for technical assis- To index cystic disease severity, we calculated the total number tance. of cysts using four histologic sections from each kidney. De- tailed procedures for histology, IF, and IHC were published previously.31 For confocal microscopy, antibody-stained im- DISCLOSURES ages were collected as Z-series sections using a Zeiss LSM 510 None confocal microscope system with ϫ40, ϫ63, and ϫ100 oil ob- jectives. For scanning electron microscopy, mice were perfused as described previously.31 The kidneys were removed, sec- REFERENCES tioned longitudinally, washed by 1ϫ PBS three times, and fixed in 2.5% paraformaldehyde. After an ethanol dehydration se- 1. Igarashi P, Somlo S: Genetics and pathogenesis of polycystic kidney ries, the samples were critical-point dried and sputter-coated disease. J Am Soc Nephrol 13: 2384–2398, 2002 with 40% gold/60% palladium microparticles to a thickness of 2. Wilson P: Polycystic kidney disease. N Engl J Med 350: 151–164, 2004 15 to 17 mm. Images of the samples were obtained using an 3. Torres VE, Harris PC: Mechanisms of disease: Autosomal dominant Electroscan E3 Environmental Scanning Electron Microscope. and recessive polycystic kidney diseases. Nat Clin Pract Nephrol 2: 40–55, quiz 55, 2006 4. Mochizuki T, Wu G, Hayashi T, Xenophontos SL, Veldhuisen B, Saris Cell Lines and Mouse Kidney Primary Epithelial JJ, Reynolds DM, Cai Y, Gabow PA, Pierides A, Kimberling WJ, Cultures Breuning MH, Deltas CC, Peters DJ, Somlo S: PKD2, a gene for All cell lines used in this study were cultured under previously polycystic kidney disease that encodes an integral membrane protein. Science 272: 1339–1342, 1996 39 described conditions. To generate primary cultures of renal 5. Sutters M, Germino GG: Autosomal dominant polycystic kidney dis- epithelia, we removed kidneys from 2-mo-old WT or Pkhd1 ease: Molecular genetics and pathophysiology. J Lab Clin Med 141: mutant mice and minced them finely with a scalpel. The 91–101, 2003 minced tissue was incubated with 0.5% collagenase type IV at 6. Montell C, Birnbaumer L, Flockerzi V: The TRP channels, a remarkably 37°C for 45 min and pipetted vigorously. The undigested tissue functional family. Cell 108: 595–598, 2002 ␮ 7. Qamar S, Vadivelu M, Sandford R: TRP channels and kidney disease: was removed by filtration through a 40- mesh filter. The re- Lessons from polycystic kidney disease. Biochem Soc Trans 35: 124– maining single cells and small organoids were washed three 128, 2007 times with PBS containing 5 mM glucose. The cells were incu- 8. The polycystic kidney disease 1 gene encodes a 14 kb transcript and bated with 10 ␮g/ml biotinylated Dolichos biflorus agglutinin lies within a duplicated region on chromosome 16. European Polycys- (Vector; B-1035) at 4°C for 60 min. Then the cells were washed tic Kidney Disease Consortium. Cell 77: 881–894, 1994 ␮ 9. Qian F, Germino FJ, Cai Y, Zhang X, Somlo S, Germino GG: PKD1 again with PBS before incubation with 50 l of CELLectin interacts with PKD2 through a probable coiled-coil domain. Nat Genet Biotin binder Dynabeads (Invitrogen) at 4°C for 30 min. Be- 16: 179–183, 1997 cause Dynabeads are superparamagnetic polystyrene beads, 10. Nauli SM, Alenghat FJ, Luo Y, Williams E, Vassilev P, Li X, Elia AE, Lu the incubated mixtures were washed twice with PBS contain- W, Brown EM, Quinn SJ, Ingber DE, Zhou J: Polycystins 1 and 2 ing 5 mM glucose using a magnetic rack. The cells were eluted mediate mechanosensation in the primary cilium of kidney cells. Nat Genet 33: 129–137, 2003 with release buffer and were plated on 24-well dishes with 10% 11. Koulen P, Cai Y, Geng L, Maeda Y, Nishimura S, Witzgall R, Ehrlich BE, FCS DMEM under 5% CO2 at 37°C. Somlo S: Polycystin-2 is an intracellular calcium release channel. Nat

466 Journal of the American Society of Nephrology J Am Soc Nephrol 19: 455–468, 2008 www.jasn.org BASIC RESEARCH

Cell Biol 4: 191–197, 2002 29. Masyuk TV, Huang BQ, Ward CJ, Masyuk AI, Yuan D, Splinter PL, 12. Hanaoka K, Qian F, Boletta A, Bhunia AK, Piontek K, Tsiokas L, Punyashthiti R, Ritman EL, Torres VE, Harris PC, LaRusso NF: Defects Sukhatme VP, Guggino WB, Germino GG: Co-assembly of polycys- in cholangiocyte fibrocystin expression and ciliary structure in the PCK tin-1 and -2 produces unique cation-permeable currents. Nature 408: rat. Gastroenterology 125: 1303–1310, 2003 990–994, 2000 30. Menezes LF, Cai Y, Nagasawa Y, Silva AM, Watkins ML, Da Silva AM, 13. Guay-Woodford LM: Autosomal recessive disease: Clinical and ge- Somlo S, Guay-Woodford LM, Germino GG, Onuchic LF: Polyductin, the netic profiles. In: Polycystic Kidney Disease, edited by Torres V, PKHD1 gene product, comprises isoforms expressed in plasma mem- Watson M, Oxford, Oxford University Press, 1995, pp 237–267 brane, primary cilium, and cytoplasm. Kidney Int 66: 1345–1355, 2004 14. Zerres K, Mucher G, Becker J, Steinkamm C, Rudnik-Schoneborn S, 31. Zhang MZ, Mai W, Li C, Cho SY, Hao C, Moeckel G, Zhao R, Kim I, Heikkila P, Rapola J, Salonen R, Germino GG, Onuchic L, Somlo S, Wang J, Xiong H, Wang H, Sato Y, Wu Y, Nakanuma Y, Lilova M, Pei Avner ED, Harman LA, Stockwin JM, Guay-Woodford LM: Prenatal Y, Harris RC, Li S, Coffey RJ, Sun L, Wu D, Chen XZ, Breyer MD, Zhao diagnosis of autosomal recessive polycystic kidney disease (ARPKD): ZJ, McKanna JA, Wu G: PKHD1 protein encoded by the gene for Molecular genetics, clinical experience, and fetal morphology. Am J autosomal recessive polycystic kidney disease associates with basal Med Genet 76: 137–144, 1998 bodies and primary cilia in renal epithelial cells. Proc Natl Acad Sci 15. Zerres K, Rudnik-Schoneborn S, Senderek J, Eggermann T, Bergmann USA101: 2311–2316, 2004 C: Autosomal recessive polycystic kidney disease (ARPKD). J Nephrol 32. Wu GQ, D’Agati V, Cai Y, Markowitz G, Park JH, Reynolds DM, Maeda 16: 453–458, 2003 Y, Le TC, Hou H Jr, Kucherlapati R, Edelmann W, Somlo S: Somatic 16. Capisonda R, Phan V, Traubuci J, Daneman A, Balfe JW, Guay-Wood- inactivation of Pkd2 results in polycystic kidney disease. Cell 93: ford LM: Autosomal recessive polycystic kidney disease: Outcomes 177–188, 1998 from a single-center experience. Pediatr Nephrol 18: 119–126, 2003 33. Wu GQ, Markowitz GS, Li L, D’Agati VD, Factor SM, Geng L, Tibara S, 17. Guay-Woodford LM, Desmond RA: Autosomal recessive polycystic Tuchman J, Cai Y, Park JH, van Adelsberg J, Hou H Jr, Kucherlapati R, kidney disease: The clinical experience in North America. Pediatrics Edelmann W, Somlo S: Cardiac defects and renal failure in mice with 111: 1072–1080, 2003 targeted mutations in Pkd2. Nat Genet 24: 75–78, 2000 18. Roy S, Dillon MJ, Trompeter RS, Barratt TM: Autosomal recessive 34. Yoder BK, Tousson A, Millican L, Wu JH, Bugg CE Jr, Schafer JA, polycystic kidney disease: Long-term outcome of neonatal survivors. Balkovetz DF: Polaris, a protein disrupted in orpk mutant mice, is Pediatr Nephrol 11: 302–306, 1997 required for assembly of renal cilium. Am J Physiol Renal Physiol 282: 19. Fonck C, Chauveau D, Gagnadoux MF, Pirson Y, Grunfeld JP: Auto- F541–F552, 2002 somal recessive polycystic kidney disease in adulthood. Nephrol Dial 35. Lin F, Hiesberger T, Cordes K, Sinclair AM, Goldstein LS, Somlo S, Transplant 16: 1648–1652, 2001 Igarashi P: Kidney-specific inactivation of the KIF3A subunit of kine- 20. Neumann HP, Krumme B, van Velthoven V, Orszagh M, Zerres K: sin-II inhibits renal ciliogenesis and produces polycystic kidney dis- Multiple intracranial aneurysms in a patient with autosomal recessive ease. Proc Natl Acad Sci U S A 100: 5286–5291, 2003 polycystic kidney disease. Nephrol Dial Transplant 14: 936–939, 1999 36. Wu Y, Dai XQ, Li Q, Chen CX, Mai W, Hussain Z, Long W, Montalbetti 21. Lilova MI, Petkov DL: Intracranial aneurysms in a child with autosomal N, Li G, Glynne R, Wang S, Cantiello HF, Wu G, Chen XZ: Kinesin-2 recessive polycystic kidney disease. Pediatr Nephrol 16: 1030–1032, mediates physical and functional interactions between polycystin-2 2001 and fibrocystin. Hum Mol Genet 15: 3280–3292, 2006 22. Yonemura K, Yasuda H, Fujigaki Y, Oki Y, Hishida A: Adrenal insuffi- 37. Anyatonwu GI, Ehrlich BE: Organic cation permeation through the ciency due to isolated adrenocorticotropin deficiency complicated by channel formed by polycystin-2. J Biol Chem 280: 29488–29493, 2005 autosomal recessive polycystic kidney disease. Ren Fail 25: 485–492, 38. Koulen P, Duncan RS, Liu J, Cohen NE, Yannazzo JA, McClung N, 2003 Lockhart CL, Branden M, Buechner M: Polycystin-2 accelerates Ca2ϩ 23. Ward CJ, Hogan MC, Rossetti S, Walker D, Sneddon T, Wang X, Kubly release from intracellular stores in Caenorhabditis elegans. Cell Cal- V, Cunningham JM, Bacallao R, Ishibashi M, Milliner DS, Torres VE, cium 37: 593–601, 2005 Harris PC: The gene mutated in autosomal recessive polycystic kidney 39. Mai W, Chen D, Ding T, Kim I, Park S, Cho SY, Chu JS, Liang D, Wang disease encodes a large, receptor-like protein. Nat Genet 30: 259– N, Wu D, Li S, Zhao P, Zent R, Wu G: Inhibition of Pkhd1 impairs 269, 2002 tubulomorphogenesis of cultured IMCD cells. Mol Biol Cell 16: 4398– 24. Onuchic LF, Furu L, Nagasawa Y, Hou X, Eggermann T, Ren Z, Berg- 4409, 2005 mann C, Senderek J, Esquivel E, Zeltner R, Rudnik-Schoneborn S, 40. Hiesberger T, Gourley E, Erickson A, Koulen P, Ward CJ, Masyuk Mrug M, Sweeney W, Avner ED, Zerres K, Guay-Woodford LM, Somlo TV, Larusso NF, Harris PC, Igarashi P: Proteolytic cleavage and S, Germino GG: PKHD1, the polycystic kidney and hepatic disease 1 nuclear translocation of fibrocystin is regulated by intracellular gene, encodes a novel large protein containing multiple immunoglob- Ca2ϩ and activation of protein kinase C. J Biol Chem 281: 34357– ulin-like plexin-transcription-factor domains and parallel beta-helix 1 34364, 2006 repeats. Am J Hum Genet 70: 1305–1317, 2002 41. Garcia-Gonzalez MA, Menezes LF, Piontek KB, Kaimori J, Huso DL, 25. Xiong H, Chen Y, Yi Y, Tsuchiya K, Moeckel G, Cheung J, Liang D, Watnick T, Onuchic LF, Guay-Woodford LM, Germino GG: Genetic Tham K, Xu X, Chen XZ, Pei Y, Zhao ZJ, Wu G: A novel gene encoding interaction studies link autosomal dominant and recessive polycystic a TIG multiple domain protein is a positional candidate for autosomal kidney disease in a common pathway. Hum Mol Genet 16: 1940– recessive polycystic kidney disease. Genomics 80: 96–104, 2002 1950, 2007 26. Bergmann C, Frank V, Kupper F, Schmidt C, Senderek J, Zerres K: 42. Torres VE, Wang X, Qian Q, Somlo S, Harris PC, Gattone VH 2nd: Functional analysis of PKHD1 splicing in autosomal recessive polycys- Effective treatment of an orthologous model of autosomal dominant tic kidney disease. J Hum Genet 51: 788–793, 2006 polycystic kidney disease. Nat Med 10: 363–364, 2004 27. Ward CJ, Yuan D, Masyuk TV, Wang X, Punyashthiti R, Whelan S, 43. Wang S, Zhang J, Nauli SM, Li X, Starremans PG, Luo Y, Roberts KA, Bacallao R, Torra R, LaRusso NF, Torres VE, Harris PC: Cellular and Zhou J: Fibrocystin/polyductin, found in the same protein complex subcellular localization of the ARPKD protein; fibrocystin is expressed with polycystin-2, regulates calcium responses in kidney epithelia. Mol on primary cilia. Hum Mol Genet 12: 2703–2710, 2003 Cell Biol 27: 3241–3252, 2007 28. Wang S, Luo Y, Wilson PD, Witman GB, Zhou J: The autosomal 44. Moser M, Matthiesen S, Kirfel J, Schorle H, Bergmann C, Senderek J, recessive polycystic kidney disease protein is localized to primary cilia, Rudnik-Schoneborn S, Zerres K, Buettner R: A mouse model for cystic with concentration in the basal body area. J Am Soc Nephrol 15: biliary dysgenesis in autosomal recessive polycystic kidney disease 592–602, 2004 (ARPKD). Hepatology 41: 1113–1121, 2005

J Am Soc Nephrol 19: 455–468, 2008 FPC Regulates PC2 Expression and Function 467 BASIC RESEARCH www.jasn.org

45. Pennekamp P, Karcher C, Fischer A, Schweickert A, Skryabin B, ARPKD gene. J Am Soc Nephrol 13: 2246–2258, 2002 Horst J, Blum M, Dworniczak B: The ion channel polycystin-2 is 51. Kaimori JY, Nagasawa Y, Menezes LF, Garcia-Gonzalez MA, Deng J, required for left-right axis determination in mice. Curr Biol 12: Imai E, Onuchic LF, Guay-Woodford LM, Germino GG: Polyductin 938–943, 2002 undergoes notch-like processing and regulated release from primary 46. McGrath J, Somlo S, Makova S, Tian X, Brueckner M: Two populations cilia. Hum Mol Genet 16: 942–956, 2007 of node monocilia initiate left-right asymmetry in the mouse. Cell 114: 52. Wu G, Tian X, Cai Y, Markowitz G, D’Agati V, Park JH, Yao L, Li L, Geng 61–73, 2003 L, Zhao H, Edelmann W, Somlo S: Trans-heterozygous Pkd1 and Pkd2 47. Shannon MB, Patton BL, Harvey SJ, Miner JH: A hypomorphic muta- mutations modify expression of polycystic kidney disease. Hum Mol tion in the mouse laminin alpha5 gene causes polycystic kidney dis- Genet 11: 1845–1854, 2002 ease. J Am Soc Nephrol 17: 1913–1922, 2006 53. Li Z, Hannigan M, Mo Z, Liu B, Lu W, Wu Y, Smrcka AV, Wu G, Li L, Liu 48. Fischer E, Legue E, Doyen A, Nato F, Nicolas JF, Torres V, Yaniv M, M, Huang CK, Wu D: Directional sensing requires G beta gamma- Pontoglio M: Defective planar cell polarity in polycystic kidney dis- mediated PAK1 and PIX alpha-dependent activation of Cdc42. Cell ease. Nat Genet 38: 21–23, 2006 114: 215–227, 2003 49. Germino GG: Linking cilia to Wnts. Nat Genet 37: 455–457, 2005 50. Nagasawa Y, Matthiesen S, Onuchic LF, Hou X, Bergmann C, Esquivel E, Senderek J, Ren Z, Zeltner R, Furu L, Avner E, Moser M, Somlo S, Guay-Woodford L, Buttner R, Zerres K, Germino GG: Identification See related editorial, “ARPKD and ADPKD: First Cousins or More Distant and characterization of Pkhd1, the mouse orthologue of the human Relatives?,” on pages 416–418.

468 Journal of the American Society of Nephrology J Am Soc Nephrol 19: 455–468, 2008